JesFord Senior Data Scientist contact profile
[email protected] Experienced Data Scientist with a passion for applied machine learning and open source soft- linkedin.com/in/jesford ware. I bring a practical product driven mindset, an obsession with best practices, and a 206.446.6874 demonstrated ability to collaborate effectively across teams to achieve goals. I am a regular website conference speaker at events including PyCon, PluralsightLIVE, and VisInPractice. jesford.github.io experience GitHub 2018–current Recursion Salt Lake City, UT @jesford Senior Data Scientist 11/2019–current Data Scientist 4/2018–11/2019 tech skills • Trained, optimized, and benchmarked deep neural networks (CNNs and Python graph NNs) to solve problems in drug discovery. pandas, numpy, scipy, • Technical lead on teams that productionized the first machine learning pytorch, keras, xgboost, models at Recursion, responsible for directing a small team of data sci- scikit-learn, jupyter, entists, and providing regular updates to VPs. seaborn, matplotlib, prefect, dash, plotly • Model interpretability including saliency maps, GradCAM, DeepDream. • Pioneered use of MCMC for Bayesian parameter fitting of experimental Git & GitHub data, and developed new visualizations to inform biologists. Bash • Active Learning for improving class imbalance & minimizing model bias. SQL • Owned creation of cleaned & standardized labelsets from noisy data, en- C/C++ abling comparison of ML models across data science teams. Travis-CI • Designed & built custom Google Cloud database for storing ML models, AWS & GCP metrics, labels, and predictions; created resources for data science users. Sphinx • Led data science team to adopt software best practices for reproducible HTML/CSS research and clean production quality code.